INDEX

Explanations

so, while

The neuron fires on passages reporting numeric data—counts, dimensions, measurements, and the punctuation that frames those statistics.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ம்தான்

0.63

នូវ

0.63

ပါ

0.62

лами

0.61

 তাতে

0.57

それでも

0.55

及其

0.54

куп

0.52

그래

0.52

জ্ঞা

0.50

POSITIVE LOGITS

 while

5.34

 mientras

5.00

while

4.97

 mentre

4.94

 enquanto

4.59

 WHILE

4.56

 whereas

4.50

 sedangkan

4.41

 tandis

4.38

 بينما

4.25

Activations Density 1.297%